TheHSTGOALSSurvey · 12...

The HST GOALS Survey: Probing the Morphology and Evolution of U/LIRGs CSIRO ASTRONOMY & SPACE SCIENCE

Sebas&an Haan 26 Sep 2013, Sydney

Luminous Infra-‐Red Galaxies •  Enhanced IR luminosity (LIRGs LIR > 1011 L☉; ULIRGs: LIR > 1012 L☉ ) •  Very gas-‐ and dust-‐rich, 10 to 500x larger SFR than normal galaxies •  Comprise >50% of cosmic IR background and dominate SF ac&vity at z > 1 (Capu&+07, Magnelli+11, Berta+11) •  Represent a cri&cal phase in the evolu&on of galaxies where most of the stellar galaxies mass is building up

Morphology and Evolu&on of LIRGS | Sebas&an Haan 2 |

Magnelli+11

Normal Galaxies LIRGs ULIRGs


Lee Armus, Aaron Evans, David Sanders, Philip Appleton, Josh Barnes, Greg Bothun, Carrie Bridge, Ben Chan, Vassilis Charmandaris, Lisa Chien, Tanio Diaz-‐Santos, Michael Dopita, David Frayer, Justin Howell, Sebastian Haan, Hanae Inami, Kazushi Iwasawa, Joseph Jensen, Lisa Kewley, Dong-‐Chan Kim, Stefanie Komossa, Steven Lord, Nanyao Lu, Barry Madore, Jason Marshall, Jason Melbourne, Joseph Mazzarella, Eric Murphy, Andreea Petric, George Privon, Jeff Rich, Shobita Satyapal, Bernhard Schulz, Henrik Spoon, Sabrina Stierwalt, Eckhard Sturm, Jason Surace, Vivian U, Tatjana Vavilkin, Sylvain Veilleux, Kevin Xu

HST Spitzer Herschel

Chandra GALEX

Groundbased: JVLA Palomar CSO ALMA


The GOALS Sample

•  Complete subset of the IRAS Revised Bright Galaxy

Sample (RBGS) with 202 LIRGS and ULIRGs �

•  Combination of imaging and spectroscopic data

from Spitzer (IRAC, IRS), HST (ACS, NICMOS,

WFC3), Chandra, GALEX, and Herschel. �

HST:

•  88 most IR luminous systems (log[LIR/L¤] > 11.4) �

•  H-, I-, and B-band (1.6μm, 814nm, 435nm) �

à 50% of nuclei are dust-obscured in B-band! �

•  Predominantly mergers and interacting galaxies�

•  Resolution: 0.15 arcs (~106pc), redshift z < 0.05 �

Probing the dust-unobscured nuclear stellar structure! �

Probing the FormaFon of Nuclear Stellar Cusps


Interac&on and &dal forces: à Gas Inflow

Final coalescence of nuclei: Dense gas in center à extreme Starburst (10—1000 M⊙/yr) à Stellar Cusp is build up

Gas evtl. expelled from center à Shutdown of star forma&on Nuclear stellar cusp remains Progenitor stars form spheroid

Time

Hopkins+08

Open quesFons: •  Can we see cusp formation during the actual starburst phase

and how does it link to the nuclear properties of elliptical galaxies?

•  How much stellar mass is typically built up in nuclear cusps and what are the critical time-scales for cusp formation?

•  Is there a link between the strength of cusps in LIRGs and the current rate of star formation in these galaxies?

•  Do we see the same fraction of cusps in LIRGs as in elliptical galaxies?


ObservaFonal evidence: Some elliptical galaxies show excess light in their radial light profiles.


Near-‐IR (Evolved Stars) Cusp (radius=0.35’’, 110pc) HST

An Example of a Nuclear Stellar Cusps

RGB Image (H-‐, I-‐, B-‐band)

H-‐band, 1.6μm H-‐band, 1.6μm

Logarithmic brightness scale!


The Nuker Profile (Lauer et al. 1995): •  Double power law •  γ is the inner (cusp) slope

Nuker Profilewith Cusp

Standard Parameter Fit(No Cusp)

Transition betweeninner and outer slope(R=0.34arcsec, 110pc)

1995AJ....110.2622L

FiLng of the nuclear stellar light and cusp In prac&ce:

2-‐dimensional fit with GALFIT (Peng et al. 2010) •  Simultaneous fikng of several components

•  Fikng and subtrac&ng of central unresolved light component


Cusp Slope and Luminosity DistribuFon


0.0 0.5 1.0 1.5Cusp Slope !

0

5

10

15

20

25

N

MergerNon-Merger

6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0LCusp log[L/L!]

0

5

10

15

20

25

N

Upper Limits

H-‐band Cusp Luminosity Nuclear Slope

Resolve nuclear cusps in 76% of LIRGs

Morphology and Evolu&on of LIRGS | Sebas&an Haan 11

RelaFonship between Current Star-‐formaFon and Cusp Strength

Haan+13

11.0 11.5 12.0 12.5LIR(8–1000µm) log[L/L!]

6

7

8

9

10

11

12

L Cusp(1.6µm)log[L/L

!]

SFR

CuspMass

Increase in Cusp Mass: Above log[LIR/L⊙]=11.9: all galaxies have large cusp NIR luminosi&es. On average, five &mes larger than for lower luminosity LIRGs (log[LIR /L⊙ ∼ 11.5]).


Cusp ProperFes as FuncFon of Merger Stage

LIR Luminosity (8—1000μm) Cusp Luminosity (1.6μm) Nuclear Density (R=1kpc)

Current starburst activity is associated with the build-up of cusps due to merger process.

1 2 3 4 5Merger Stage

10.5

11.0

11.5

12.0

12.5

L IRlog[L/L !]

N = 7 N = 21 N = 14 N = 11 N = 141 2 3 4 5

10.0

10.2

10.4

10.6

10.8

11.0

11.2

11.4

L Nuclog[L/L !]

N = 6 N = 16 N = 14 N = 6 N = 111 2 3 4 5

7.5

8.0

8.5

9.0

9.5

10.0

10.5

11.0

L Cusplog[L/L !]

N = 6 N = 16 N = 14 N = 6 N = 11


Stellar Masses and Timescales for Build-‐up of Cusps

•  Increase in stellar cusp mass towards late merger stages:

ΔMcusp = (7 ± 3.5) × 109M¤ (ΥH = 0.3 ± 0.15, from STARBURST99 and

dynamical mass measurements)

•  Typical merger &mescale from mid-‐stage to late stage: ≈500 Myrs

•  Timescale to build-‐up stellar cusp mass:

Δt = (60 ± 30) Myrs (based on current SFR, ≈220 M¤ /yr, and 50% of LIR from cusp region)


0.0 0.5 1.0 1.5024681012141618

N

0.0 0.5 1.0 1.5Cusp Slope !

0

5

10

15

20

N

Comparison of Cusp DistribuFon to EllipFcals Early

Type Galaxies

(Lau

er et a

l 200

7)

LIRG

s Sta&s&cs for same range in host galaxy mass:

-‐23 < MH <-‐25.5 [mag] (98% of the LIRGs in our sample )

cusp/core = 0.7:1

cusp/core > 3.2:1

Late Stage Merger


3.3. Intermediate Galaxies

The intermediate classification appears to be highly dependenton resolution, as Rest et al. (2001) noted. Indeed, we consider allthe galaxies so classified by Rest et al. (2001) to be core galaxies,given the finer resolution limits adopted here for their WFPC2profiles. At the same time, we consider NGC 2902, NGC 4482,NGC 5796, and NGC 5898, which were classified as power lawsby Rest et al. (2001), to be intermediate galaxies. Likewise, wefind the intermediate classification in the Ravindranath et al.(2001) sample to vary with choice of resolution limit. In a sense,the tests conducted above might lead one to relabel the interme-diate galaxies with 0:3 < ! 0 < 0:5 as ‘‘ambiguous galaxies.’’ Thevalley in a bimodal distribution is a region where outliers fromeach mode coexist and classification of any given galaxy cannotbe done with any degree of reliability. It is likely that with betterresolution some of the present intermediate galaxies would beclassified as core galaxies. NGC 3585 and NGC 4709, for ex-ample, have well-defined breaks but have ! 0 of 0.31 and 0.32,respectively, falling only just above the core boundary; theywould probably fall below it with better resolution.

At the same time, we are left with the real physical question ofwhy intermediate galaxies are rare and the possibility that in atleast some cases their central structure may offer a unique win-dow on the problem of core formation.AsRest et al. (2001) noted,many intermediate galaxies have small " and do not show asstrong breaks as the core galaxies, even accounting for the higher! values. In the present sample we have difficulty identifying anygalaxy with ! 0 ! 0:4 that shows a sharp break and well-definedinner cusp: the best candidates for this appear to be NGC 2902,NGC 5796, and NGC 5898. To explore the utility of the inter-mediate classification, we choose to preserve the 13 intermediategalaxies in the sample as a separate class. This appears to be sen-sible based on the central parameter relationships presented inLauer et al. (2007); the intermediate galaxies do not appear to fallon the relationships defined by core galaxies.

4. A NONPARAMETRIC DEMONSTRATIONOF BIMODALITY IN SPACE DENSITY

Parametric models provide a means to represent data in a sim-ple way that allows one to determine underlying physical rela-tionships. However, care has been taken so that unknown biasesin the parameterization do not affect the overall results. This con-cern is especially important when studying the luminosity den-sity. Since it is calculated by deprojecting the surface brightnessprofile, any biases in the brightness profile model will be ampli-fied. Luminosity density is more fundamental than surface bright-ness, thus it is important to verify that the bimodality of cuspslopes in projection truly reflects the intrinsic distribution of spacedensity slopes. We therefore bolster the result in the previous sec-tion with a nonparametric approach.

The uncertainties for nonparametric models will generally belarger than those resulting from parameter fits and should pro-vide a more realistic estimate. The comparison between the twoapproaches was done in Gebhardt et al. (1996), where we foundthat the Nuker law provides an accurate estimate of the surfacebrightness profile in the central regions of early-type galaxies.While detailed analysis of the luminosity density, including dy-namical studies, should incorporate a nonparametric approach,use of the Nuker law facilitates comparison with theoreticalmodels and parameter correlation studies. Therefore, both areappropriate.

We follow the same procedures as in Gebhardt et al. (1996) toestimate the surface brightness and luminosity density profiles.The technique represents the measured brightness profiles withsmoothing splines and then inverts the splines into density space.Monte Carlo simulations are used to establish the errors in thefinal density profiles. As the present analysis is identical to thatof Gebhardt et al. (1996), we refer the reader to that paper for acomplete description.

Figure 6 shows the luminosity density profiles normalized inboth radius and luminosity density of all galaxies for which we

Fig. 5.—Limiting logarithmic slope ! 0 of the surface photometry profiles ofthe sample as a function of galaxy luminosity. The solid line shows the distri-bution of ! 0 for just those galaxies with "20:5 # MV > "22:5 (double verticallines) modeled as the sum of two Gaussians with equal dispersions. The dashedline shows the distribution when the two Gaussians are allowed to have differingdispersions. The normalization is arbitrary but is the same for both distributions.

Fig. 4.—Limiting logarithmic slope ! 0 of the surface photometry profiles ofthe sample as a function of distance. The solid line shows the distribution of ! 0 forjust those galaxies with D $ 50 Mpc (vertical line) modeled as the sum of twoGaussians with equal dispersions. The dashed line shows the distribution whenthe twoGaussians are allowed to have differing dispersions. The normalization isarbitrary but is the same for both distributions.

CENTERS OF EARLY-TYPE GALAXIES WITH HST. VI. 239No. 1, 2007

Lauer+07

Stellar Mass of Galaxy

Cusp

Core

Cusp-‐Core Dichotomy in Early Type Galaxies


Stellar Mass of Galaxy

Cusp

Core

Early Type Galaxies vs LIRGs

Haan+13


What’s going on?

Cusp DestrucFon: via BH binary in a subsequent dry merger (no gas) à Forma&on of massive core ellipFcals

Different Cusp-‐Galaxy FormaFon Scenario E.g. most cusp ellipFcals formed at an early phase of the universe when most galaxies were smaller and had larger gas frac&ons than in today’s LIRGs.

? Cusp

BHs

Black hole ‘scouring’

Different environment during formaFon and evoluFon: a) gas fracFon, b) mass of progenitor galaxies, c) merger density/history


Summary & Conclusions •  Measurement of the nuclear structure provides important insight into

merger and starburst history of galaxies.

•  Nuclear stellar cusps are found in at least 76% of (U)LIRGs.

•  Cusp strength and luminosity increase with merger stage and total IR luminosity (excluding AGN), confirming models that recent starburst ac&vity is associated with the build-‐up of cusps.

•  Nuclear stellar structure becomes more compact towards late merger stages.

•  Comparison to local early-‐type galaxies: a) Local (U)LIRGs have a significantly larger cusp frac&on b) Most LIRGs have host galaxy luminosi&es (H-‐band) similar to core ellip&cals which is roughly one order in magnitude larger than for cusp ellip&cals.

CSIRO ASTRONOMY & SPACE SCIENCE

Thank you CSIRO Astronomy & Space Science SebasFan Haan Postdoc Fellow E sebas&[email protected] w www.atnf.csiro.au/people/Sebas&an.Haan/

TheHSTGOALSSurvey · 12...

Documents

Transcript of TheHSTGOALSSurvey · 12...